Really looking forward to presenting our paper "HuManiFlow: Ancestor-Conditioned Normalising Flows on SO(3) Manifolds for Human Pose and Shape Distribution Estimation" at CVPR 2023! w/ Ignas Budvytis and Roberto Cipolla Paper + code: github.com/akashsengupta1… 1/3

Flight canceled. First time unable to attend CVPR due to visa issues since 2012. Being part of the organization team, I've seen the effort made by the organizers to assist with visa processes. The frustration is shared by everyone involved. IRCC is not doing any favor for Canada.

What to Learn: Features, Image Transformations, or Both? Yuxuan Chen, Binbin Xu, Frederike Dümbgen, Timothy D. Barfoot tl;dr: when matching day to night, first do img2img night2day, then learn the feature extractor, which works on real and generated day arxiv.org/abs/2306.13040

very interesting work. similar idea to our CVPR2023 paper IMP: Iterative Matching and Pose Estimation with Adaptive Pooling w/ Ignas Budvytis Roberto Cipolla arxiv.org/pdf/2304.14837… Source code including the training part is available at github.com/feixue94/imp-r…

Moon bridge Yongqing Fang, Guangzhou, China [📷 Yier Wang]

SimpleNeRF: Regularizing Sparse Input Neural Radiance Fields with Simpler Solutions Nagabhushan Somraj, Adithyan Karanayil, Rajiv Soundararajan tl;dr: augmented NeRF models -> depth estimation -> supervision arxiv.org/pdf/2309.03955…

Dynamic Mesh-Aware Radiance Fields paper page: huggingface.co/papers/2309.04… Embedding polygonal mesh assets within photorealistic Neural Radience Fields (NeRF) volumes, such that they can be rendered and their dynamics simulated in a physically consistent manner with the NeRF, is

Introducing 𝗥𝗧-𝗫: a generalist AI model to help advance how robots can learn new skills. 🤖 To train it, we partnered with 33 academic labs across the world to build a new dataset with experiences gained from 22 different robot types. Find out more: dpmd.ai/TW_RT-X

what's the future of nerf? better accuracy or higher efficiency?

Excited for #ICCV2023 in person! We invite you to visit our poster of ‘Active Neural Mapping’ at Foyer Sud 109 tomorrow (5th Oct. 10:30 AM - 12:30 PM). TLDR: Weight perturbation on the INR parameters can guide the mobile agent to explore the unknown environment on-the-fly.

🚨 BREAKING: OpenAI just released 128k context GPT-4 Turbo at 1/3rd the price. This means GPT-4 is cheaper, faster, and can now read a prompt as long as an entire book. The announcement comes with additional huge news for developers: - New Assistants API for Code Interpreter

FocusTune: Tuning Visual Localization through Focus-Guided Sampling Son Tung Nguyen, Alejandro Fontan, Michael Milford, FTSE 🤖🚘🐀🧠📘🗣️, Tobias Fischer tl;dr: reproject 3D coordinates onto 2D image plane and sample within a local neighborhood of the reprojected points arxiv.org/pdf/2311.02872…

Had great experience at ICVSS 2023. Looking forward to attending ICVSS 2024.

Excited to be presenting “DiffHuman: Probabilistic Photorealistic 3D Reconstruction of Humans” at #CVPR2024! Project page: akashsengupta1997.github.io/diffhuman/ Huge thanks to my awesome Google AI internship hosts: Thiemo Alldieck, Nikos Kolotouros, Enric Corona, Andrei Zanfir, Cristian Sminchisescu! 1/n

MATCHA:Towards Matching Anything feixue, Sven Elflein, Laura Leal-Taixe, Qunjie Zhou tl;dr: diffusion model->semantic+geometric features->transformer-based fusion->enhanced diffusion features->w/ DINOv2->unified feature->geometric/semantic/temporal matching arxiv.org/abs/2501.14945